ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Feb. 1972, p. 116-122Copyright @ 1972 American Society for Microbiology

Vol. 1, No. 2Printed in U.S.A.

Influence of Different Media and Bloods on theResults of Diffusion Antibiotic

Susceptibility TestsV. C. BRENNER' AND J. C. SHERRIS

Department of Microbiology, School of Medicine, University of Washington, Seattle, Washington 98195

Received for publication 30 September 1971

As part of an International Collaborative Study, the influence of medium andcertain medium components on the results of disc susceptibility testing with nineantibiotics was examined and statistically evaluated. Four basic media, TrypticaseSoy Agar, Grove and Randall's formula # 9 agar, and Mueller-Hinton agar with andwithout 1 mg of L-tryptophan per 100 ml, were used with six bacterial strains. Signifi-cant differences in zone diameters occurred with most antibiotics. The largest con-

sistent differences were seen with tetracycline, and appeared to result from varyingdegrees of chelation of the antibiotic with free cations in the media. Reproduci-bility studies on different batches of Mueller-Hinton agar from two manufacturersshowed some statistically significant differences, which were small except withtetracycline. Reproducibility between the products of a single manufacturer wasexcellent. The results of experiments to determine the effect of the addition of 5%blood of different species to agar medium showed that medium containing citratedor defibrinated horse, rabbit, sheep, or human blood yielded closely similar zonediameters, except that tetracycline zones were larger with citrated blood. A singletype of medium should be selected for routine susceptibility tests whenever possible,and reproducibility of performance of the products of different manufacturersshould be sought.

During the initial stages of an InternationalCollaborative Study on antibiotic susceptibilitytesting techniques, experiments were carried outto determine the reproducibility of results fromdifferent laboratories when the same strains,reagents, media, and methods were used, and alsoto compare results on the same strain obtained bythe routine methods used in the different labora-tories.

These experiments indicated a number of actualand potential sources of variation that influencethe results of susceptibility tests, and severalparticipating laboratories undertook further in-vestigations to define the significance and extentof these variations. The studies reported herewere concerned with the reproducibility of resultson different media and on different batches ofmedium. These variables appeared to be impor-tant contributors to differences in results betweenlaboratories and bore directly on the selection of amedium for any proposed reference method. Thestudy also included a statistical comparison ofresults obtained when blood-agar prepared with

I Present address: Clinical Center, National Institutes of Health,Bethesda, Md. 20014.

defibrinated and citrated bloods from severalmammalian species was used. This informationwas pertinent since supplementation was requiredfor satisfactory growth of some organisms withthe media examined, and because there are geo-graphic differences in the availability of differentanimal bloods.The results of these studies have been described

in abstract in the Report of the InternationalCollaborative Study (Acta Pathol. Microbiol.Scand. Sect. B, Suppl. 217, 1971). They are pre-sented here in detail with the technical and statis-tical methods employed.

MATERIALS AND METHODSMicroorganisms. The organisms used were clinical

isolates obtained from patients in the UniversityHospital, University of Washington.

Antibiotic discs. Discs containing the antibioticslisted in the tables were used. All except those con-taining cephalothin were provided by Difco Labora-tories. Cephalothin discs were obtained from EliLilly & Co. Discs were from Food and Drug Adminis-tration certified batches and were of the antibioticcontents specified by Bauer et al. (2).

116

on June 25, 2018 by guesthttp://aac.asm

.org/D

ownloaded from

INFLUENCE OF MEDIA ON SUSCEPTIBILITY TESTS

Media. The following basic media were used alongwith certain variations described in Results.

Trypticase Soy Agar (TSA) contained 15 g of apancreatic digest of casein (Trypticase, BBL), 5 g of apapaic digest of soya meal USP (Phytone, BBL), 5 g ofNaCl, and 15 g of agar in 1 liter of distilled water. ThepH of the medium was adjusted to 7.3.

Grove and Randall's (6) formula # 9 (GR) mediumcontained 17 g of a pancreatic digest of casein USP,3 g of a papaic digest of soya meal USP, 5 g of NaCl,2.5 g of dipotassium phosphate, and 15 g of agar in 1liter of distilled water. The pH was adjusted to 7.3.

Mueller-Hinton (MH) agar contained dehydratedinfusion from 300 g of beef, 17.5 g of an acid hydrol-ysate of casein, 1.5 g of starch, and 17 g of agar in afinal volume of 1 liter of distilled water. The pH wasadjusted to 7.4.

Unless otherwise specified, dehydrated commercialpreparations of TSA and MH agar were used. MHagar was obtained from BBL and from Difco Lab-oratories. GR medium was prepared in the labora-tory according to the above formula. The broths ofeach of these media were prepared in the laboratoryand had the same constitution as their correspondingsolid media except for the exclusion of agar. Whenblood-agar was used, it was prepared by adding 5%10blood (final concentration) to TSA.

Bloods. Defibrinated and citrated horse, rabbit,sheep, and human bloods were obtained. The citratedbloods contained a final concentration of 1.25%',,'sodium citrate, which was that concentration beingused by the commercial enterprise supplying theanimal bloods. Human blood was obtained fromvolunteers. All bloods were collected within 1 day ofeach other, refrigerated, and used within 1 week.

Disc diffusion susceptibility testing technique. Themethod used was based on that described by Ericssonet al. (4). Amounts of 60 ml of medium were used in14-cm petri dishes, which gave a depth of approxi-mately 4 mm. The inoculum was prepared by growingthe test organism overnight in 5 ml of GR broth anddiluting this in 0.9% sodium chloride solution. A10-2 dilution was used for staphylococci and entero-cocci, and I0-3 for Escherichia coli, Klebsiella species,and Proteus species. Plates were seeded by flooding 2ml of the inoculum onto the surface and drawing offthe excess fluid with a pipette. The surfaces were thenallowed to dry at room temperature, with the lidstipped, for 10 min before the application of discs.After this drying procedure, the surfaces of the agarhad no visible film of fluid. A maximum of nine differ-ent antibiotic discs were placed on each plate andpressed onto the agar to insure contact. After placingthe discs, a prediffusion period of 3 hr at room tem-perature was allowed before overnight incubation at37 C. Inhibition zone diameters were measured withcalipers. The end point was taken as an abrupt declineof growth. The plates were always coded and ran-domized before being read, and were read by twopersons.

Broth dilution method. Twofold dilutions of theantibiotic were prepared in broth. The final volume ineach tube was 2 ml. Each tube was inoculated with 2drops of the inoculating suspension delivered with a

pipette calibrated to deliver 30 drops per ml. Theinoculating suspension was prepared by diluting anovernight broth culture to the MacFarland # 1 bariumsulfate standard (12). An uninoculated and inoculatedcontrol of the medium without antibiotic was alsoprepared. Tubes were incubated overnight at 37 C.The minimal inhibitory concentration (MIC) wastaken as the lowest concentration of antibiotic whichinhibited visible growth.

Statistical analyses. In the comparative studies to bedescribed, four replicate plates of each medium wereused for each organism unless stated otherwise. Singlediscs of each of the antibiotics under examination wererepresented on each plate. An analysis of variance wasperformed by use of the variance ratio, F, whenevermore than two sample means needed to be compared.If a significant difference was found, the Student rangetest was applied to determine k, the minimum signifi-cant difference in zone diameters in millimeters.Statistically significant differences greater than 2 mmwere considered to be of potential practical signifi-cance. Subsidiary investigations showed that differ-ences in readings between the two readers rarelyexceeded this figure.

RESULTSInfluence of different media on inhibition zone

diameters. Comparisons were made of the resultsof diffusion tests performed with TSA, GRmedium, MH agar, and MH agar + 1 mg ofL-tryptophan per 100 ml (MH + T). The firstthree media were examined as possible referencemedia for continuation in the InternationalCollaborative Study; MH with added tryptophanwas included because acid hydrolysates of caseinare known to be deficient in this amino acid andbecause a supplemented medium has been used inJapan for many years (10). Two strains of Staphy-lococcus aureus and one strain each of Klebsiellasp., E. coli, P. mirabilis, and an enterococcus weretested. Each strain was examined simultaneouslyon the four different media, with four replicateplates for each medium. Table 1 summarizes theresults, and shows that there were significantdifferences between zone diameters on at least twoof the media with most antibiotics and organisms.The major statistically significant differences aredescribed below, and a detailed example of thetype of results obtained is given in Table 2.

Methicillin and ampicillin zone diameters werelarger on MH + T and MH agar than on GRmedium and TSA with all organisms tested.Similarly, with penicillin, cephalothin, strepto-mycin, and kanamycin, zones on MH + T andMH agar were generally larger for all organisms,with the following exceptions. (i) The entero-coccus produced consistently larger zones on GRmedium and TSA for these antibiotics, and (ii)S. aureus 2 had largest zones on MH + T andsmallest zones on MH agar with penicillin and

117VOL. 1, 1972

on June 25, 2018 by guesthttp://aac.asm

.org/D

ownloaded from

BRENNER AND SHERRIS ANTIMICROB. AG. CHEMOTHER.

TABLE 1. Magnitude of statistically significant differences (in millimeters) in mean zonediameters obtained on four different mediaa

S. aureus 2

4.2(MH + T >MH)

1.8(TSA > MH)

1.6(TSA > GR)

5.8(MH + T >GR)

4.5(MH + T >MH)

7.4(MH + T >GR)

E. coli

NSC

NS

NS

1.7(MH & MH +T > GR)

3.5(GR > MH +T)

2.8(MH & MH +T > GR)

7.8(GR > TSA)

Klebsiella

NS

1.5(MH > TSA)

NS

3.2(MH + T >GR)

2.5(GR > MH +T & MH)

5.0(MH + T >GR)

6.7(GR > TSA)

Proteuts

3.8(MH + T >GR)d

NS

4.6(MH + T >GR)

3.6M(H + T >GR)

6.0(MH + T >GR)

Enterococcus

1.8(GR & TSA >MH +T &MH)

1.7(TSA > MH +T & MH)

NS

3.4(GR & TSA >MH + T)

2.2(GR & TSA >MH + T)

7.2(GR > TSA)

a Trypticase Soy Agar (TSA), Grove and Randall's forrnula#9 (GR), Mueller-Hinton agar (MH), and MH plus 1 mg of L-

tryptophan/100 ml (MH + T).b Not tested.c No statistically significant differences in zone sizes on the four media..I The average difference between the two inedia showing the largest difference in zone sizes. The differences noted are all statis-

tically significant (P = 0.05).

TABLE 2. Example of meanti zonie diameters inmillimeters of S. aurents I oni four media

MediaAntibiotic kaiue'

TSA GR Ml rMVLH+T

Penicillin ....... 42.2 40.5 43.2 45.0 1.2Cephalothin. 40.2 40.0 40.8 43.2 1.2Tetracycline. 21.2 28.2 25.5 25.8 1.1Chloramphen-

icol ........... 30.2 28.8 30.2 30.5 l.3Erythromycin... 34.2 33.0 33.0 35.5 1.2Streptomycin.... 19.8 18.8 26.2 26.2 1.6Kanamycin. .... 28.0 25.8 30.2 30.5 0.9

a Minimum difference in millimeters betweenany two media required for statistical significance.

cephalothin. With chloramphenicol, zones werelargest on TSA except for those with P. mirabilis,which averaged 4.6 mm larger on MH + T thanon GR medium. All organisms tested behavedsimilarly with tetracycline in that zones were con-sistently 4 mm or more smaller on TSA than on

the other three media. With polymyxin B, zonesfor both E. coli and Klebsiella were larger on GRand TSA than on the two MH media.These experiments thus showed systematic

differences in the effects of the different mediawith some of the antibiotics, but in other cases theresults reflected strain differences. These differ-ences could not be accounted for by the slightdifferences in pH between the media (TSA, pH7.3; GR medium, pH 7.2 to 7.4; MH agar andMH + T, pH 7.4), or by differences in theirgrowth-supporting activities. Except for the entero-coccus, which gave equivalent growth on allmedia, growth was more luxuriant on MH + Tand MH agar than on GR medium and TSA.With most antibiotics, zone edges were morediffuse and more difficult to measure on MH + Tthan on the other media.

Specific influence of medium on tetracyclinezones. In the experiments described above, consist-ently and significantly larger tetracycline zoneswere obtained on MH agar and GR medium thanon TSA. GR medium and TSA differed mainly in

118

Antibiotic

Ampicillin .....

Cephalothin ...

Chloramphen-icol .......

Erythro-mycin ... ..

Kanamycin ....

Penicillin .. .

Polymnyxin B..

Streptomycin .

Tetracycline.

S. autreuis t

_b

3.2(MH + T >GR)

1.4(TSA > GR)

2.5(MH + T >GR & MH)

4.7(MH + T >GR)

4.5(MH + T >GR)

7.4(MH & MH +T > GR)

7.0(GR > TSA)

on June 25, 2018 by guesthttp://aac.asm

.org/D

ownloaded from

INFLUENCE OF MEDIA ON SUSCEPTIBILITY TESTS

that the formulation of GR medium includesdipotassium phosphate, whereas that ofTSA doesnot. This reagent was therefore added to TSA inequivalent concentrations. Tetracycline zonediameters for S. aureus on the modified mediumwere comparable to those on GR and MH media.These observations suggested that TSA containeda substance which interfered with tetracyclineactivity and which was counteracted by phos-phate. It is known that loss of tetracycline activityoccurs by chelation of the drug with divalent andmultivalent cations (15), and tests were made todetermine whether this could account for themedium-associated differences (Table 3).

Broth dilution and agar diffusion tests weremade with the use of TS and MH media supple-mented with calcium, magnesium, and ferrouschlorides, sodium citrate, ammonium oxalate,and dipotassium phosphate. MIC values wereincreased and zone diameters were decreased byaddition of the divalent cations, whereas thereverse was encountered when citrate, oxalate, orphosphate was added. Zone sizes for cephalothinand chloramphenicol were unaffected, thus ex-cluding nonspecific effects on zone diameters.These results indicated that differences in freecation concentrations could cause differences inzone diameters and MIC values of the magnitudeseen with different media.Medium components were interchanged be-

tween TSA and MH agar to determine the sourceof the antagonizing substance in TSA. The resultsof interest are shown in Table 4. They indicatethat the substance was present in Trypticase, apancreatic digest of casein. Trypticase had nosuch effect in the presence of an acid digest ofcasein, indicating that the latter contained aneutralizing factor.

TABLE 3. Effect of certain reagents on the activityof tetracycline against S. aureus in Trypticase

Soy and Mueller-Hinton media

Trypticase Soy Mueller-Hinton

SupplementMIC Zone MIC Zone(,ug/mI) (mmn) (pg/mi) (mm)

None .................. 5.0 23 1.25 30CaCi,, 5 X 103 M ...... 5.0 23 2.5 22MgCl2, 5 X 10 M.. 10.0. 20 2.5 25FeC1,2 5X1 M.. 20.0 11 2.5 16K2HPO4, 5 X 1r3 Mm... 2.5 24 1.25 31K2HPO4, 1.6 X 1-2 Ma 1.25 29Na3C.H507, 5 X 10-3 M 0.63 29 0.63 33(NH4) 2C204, 5 X 103 M 2.5 29 1.25 28

a Amount of K2HPO4 in Grove and Randall'sformula.

TABLE 4. Effect of interchanging media componentson tetracycline zone diameters ofS. aureus

Medium Avg. zone diam(mm)

Mueller-Hinton agar (MH) ......... 26Trypticase Soy Agar (TSA) ......... 20MH with Trypticase substituted forCaa............................ 19

TSA with Ca substituted forTrypticase........................ 25

MH with both Ca andTrypticase........................ 26

TSA with both Ca andTrypticase........................ 26

a Acid hydrolysate of casein.

Batch to batch reproducibility ofNH agar. It isimportant for any well-standardized laboratoryprocedure to utilize media or reagents which haveminimal variability in performance. For thisreason it was necessary to determine the degree ofbatch to batch variability that a susceptibilitytesting medium would show. The reproducibilityof 14 batches ofGR was first tested. Two batcheswere of commercial dehydrated medium and 12were prepared from combinations of four batchesof pancreatic digest of casein USP and three ofpapaic digest of soy meal. Tests with a strain ofS. aureus yielded largest mean zone diameterdifferences of between 2.5 and 5.3 mm with six ofeight antibiotics tested. Our attention thus turnedto the reproducibility of different batches of MHagar.

In an initial investigation, four batches of MHagar, two from each manufacturer, were tested. Afully sensitive strain of S. aureus was tested onfive replicate plates from each batch with peni-cillin, methicillin, cephalothin, streptomycin,kanamycin, erythromycin, and tetracycline. Thevariance ratio, F, was used to detect significantdifferences between any of the four batches foreach antibiotic. Significant differences were de-tected among the four batches with kanamycin,streptomycin, tetracycline, and cephalothin, butwere of less than 2 mm except with tetracycline.With tetracycline, there was an average differenceof 3.5 mm between the products of the two manu-facturers. There was, however, good agreementbetween the two batches of each manufacturer.

In a second study, six additional batches ofMHagar were tested, comprising three from eachmanufacturer. One strain of S. aureus and one ofE. coli were tested. The findings with S. aureuswere similar to those found in the initial study,but with E. coli the variability was greater. Theresults of this comparison are given in Table 5,where a number of significant differences can be

VOL. 1, 1972 119

on June 25, 2018 by guesthttp://aac.asm

.org/D

ownloaded from

BRENNER AND SHERRIS

TABLE 5. Differenices in zoiie diameters oni sixbatches of Mueller-Hinltonz cagar

Antibiotic

Penicillin.....

Methicillin. ..

Cephalothin.Streptomycin.Kanamycin ..Erythro-mycin ......

Tetracycline.Ampicillin....

PolymyxinB.........

S. aureits

1.5

0.9

1.30.70.9

1 .0

1.2Notdone

Notdone

lmdh

1.9

1.7

1.81.60.9

0.9

4.0.Not

done

Notdone

E. coli

k lmd

Not Notdone done

Not Notdone done1.3 2.41.2 1.60.9 1.9

Notdone1.22.0

0.7

Notdone5.62.5

2.4

a The k value represents the minimum differ-ence in zone diameters (in millimeters) betweenany two batches required for statistical signifi-cance.

hThe largest average difference that occurredamong the six batches.

TABLE 6. Differenices in zonie dianmeters betweetbatches of Mueller-Hiintol agar fronm maiw-

facctuirers A anztd B

Organism and drug A B AanB(3 batches) (3 batches) (6 batches)

S. alureisTetracycline..... 3.2t 1.2 4.0

E. coliCephalothin 0.5 1.0 2.4Tetracycline 2.8 1.0 5.6Ampicillin. 1.0 0.9 2.5PolymyxinB 0.9 1.5 2.4

a The largest mean difference between zonediameters on different batches.

observed. The five differences which were greaterthan 2 mm were looked at further by comparingthe reproducibility of the batches from eachmanufacturer separately. This information isshown in Table 6, and it can be seen that, with theexception of tetracycline, batches from a singlemanufacturer showed good reproducibility. Withtetracycline, results were essentially identicalamong the batches from manufacturer B, butshowed greater variation among those frommanufacturer A.

Influence of four species of defibrinated andcitrated blood on inhibition zone diameters. The

extent to which the incorporation of blood intoa test medium would affect zone sizes was a fac-tor of importance, since it was recognized that theaddition of blood is necessary for growth ofcertain organisms and since different species ofblood are utilized in different locales.

TABLE 7. Effect of blood onI zon2e diamiieter-s

Citratedvs. Amn

Antibiotic and defibrinated' bloodsborganism

Signif- lmd5 Signif-I lmdicance m icanice!

AmpicillinE. coli. NSeKlebsiella... NSS. aurCeus... NS

CephalothinE. coli. NSKlebsiella(.. NSS. ainrelis... NS

Cloramphen i-colE. coli. NSKlebsiella... NSS. aiurelus. NS

ErythromycinS. aulreus. NS

KanamycinE. coli. ...NSKlebsiell/i... NSS. aturels. NS

PenicillinS. alureus.. NS

Polymyxin BE. coli. *Klebsiella... **

StreptomycinE. coli. NSKlebsiellc... NSS. autreius. NS

TetracyclineE. coi. **Klebsiella.. **S. auireus. **

- NS**f

--- NS

NS_*.,_ **

- NSi**

NS

- NS-- NS

**

- NS

0.6 NS0.7 *

-- NS_**

3.2 NS3.8 NS4.7 NS

I _.

I1.9

Blood vs.no bloode

§Signif- lmdicance

NSNSNS

NS0.7 **1.9 1**

NS1.1 **

NS

0.8 **

- NS_ NS

0.7 NS

- NS

- NS0.6 **

1.0 NSNS

1.5 **

_**

- **

**

1.22.0

1.4

1.5

0.8

1.9

4.75.13.6

a Comparison of zones on citrated blood-agarand the respective zones on defibrinated blood-agar for all species of blood used.

I Comparison of zones on four types of blood-agar: horse, human, sheep, and rabbit.

c Comparison of zones on medium withoutblood to the respective zones with any of the fourspecies of blood incorporated.

d The largest mean differences in millimeters.e There were no statistically significant differ-

ences detected.f Statistically significant differences were ob-

served (P = 0.01).9 Statistically significant differences were ob-

served (P = 0.05).

120 ANTIMICROB. AG. CHEMOTHER.

on June 25, 2018 by guesthttp://aac.asm

.org/D

ownloaded from

INFLUENCE OF MEDIA ON SUSCEPTIBILITY TESTS

For these experiments, TSA plates were pre-pared with and without the addition of citratedor defibrinated human, horse, sheep, and rabbitblood. Single strains of E. coli, Klebsiella sp., andpenicillin-susceptible S. aureus were tested on allmedia against the appropriate antibiotics.

Duplicate plates for each type of blood-agarwere used for each organism. Statistical compari-sons were made between results on citrated anddefibrinated blood-agars, among results on thefour species of blood, and also between TSA withand without blood. All comparisons yieldingsignificant results, except those with tetracycline,were then retested with four replicates. The com-bined analyses are shown in Table 7. Data onsignificant results (except with tetracycline) arederived from the larger experiment.

In the comparison between citrated and de-fibrinated blood-agars, there was no significantdifference in zone sizes on TSA containing citratedblood as opposed to defibrinated blood (regard-less of species) except with polymyxin B andtetracycline. The differences in polymyxin zonesizes were small (0.6 to 0.7 mm) and cannot beconsidered of practical importance. For tetra-cycline, however, zones on citrated blood-agarwere considerably larger than on defibrinatedblood-agar for each of the species of blood. Thisdifference can be explained by the effect of citrate

TABLE 8. Inhibition zone diameters of a strain ofS. aureus tested onz agar conitainiing four

different species of blood

Antibiotic

AmpicillinCephalothin

ChloramphenicolErythromycin

Kanamycin

PenicillinStreptomycin

Tetracycline

ka (mm)

0.7

0.6

0.7

0.5

Blood

HumanHorseSheepRabbit

SheepRabbitHumanHorseHorseRabbitSheepHuman

RabbitHorseSheepHuman

in counteracting inhibition of tetracycline activitycaused by chelation with certain cations.To compare the results among the four species

of blood, both citrated and defibrinated bloodagars of each species were included together inthe analysis, except for tetracycline and poly-myxin B which showed significant differencesbetween citrated and defibrinated blood. Thecomparison showed a number of statisticallysignificant differences in zone sizes on the differentblood-agars, but all of these differences were 2mm or less. Table 8 is an example of how thedifferent bloods compared in actual zonediameters.Most of the antibiotics tested showed little

change in zone size (< 2 mm) with the addition ofeither citrated or defibrinated blood as comparedwith TSA without blood. The exception to thiswas tetracycline, with which significant differ-ences were again attributable to the influence ofcitrated blood.

DISCUSSIONThe experiments described here showed that

significant, and sometimes large, variations inantibiotic inhibition zone diameters may be en-countered when different media are used fordiffusion susceptibility tests. The relative order ofzone diameters on the different media varied fromantibiotic to antibiotic and in some cases with thebacterial species being tested. These results couldnot be explained by differences in growth rates, inagar content, or in initial pH. In the case of tetra-cycline, the activity of the antibiotic in diffusiontests was shown to be associated with variation inthe concentration of free divalent cations inmedia, and thus with the extent of chelation of theantibiotic. This is believed to explain the differ-ences encountered.The majority of earlier studies by others led to

similar conclusions, but for the most part theywere not statistically evaluated. Bauer (1), Bulgerand Nielson (3), Guthrie (7), Neter et al. (13),Trainer (14), and Kanazawa(l 1) all reported con-siderable influence of different media or mediumcomponents on the results of susceptibility tests. Incontrast, however, Isenberg (9) concluded from acomparison of 11 media that results were essen-tially identical on all of them. More recently,Garrod and Waterworth (5) described a strikingeffect of differences in magnesium content ofmedia and agars on the results of gentamicinsusceptibility tests with some strains of Pseudo-monas aeruginosa.The consensus of these findings strongly sup-

ports the need for selection of a single mediumformula for the majority of routine susceptibilitytests so that reproducibility of results between

a Minimum difference that is necessary in orderfor that difference to be statistically significant.

b No significant differences.

VOL. 1, 1972 121

on June 25, 2018 by guesthttp://aac.asm

.org/D

ownloaded from

122 BRENNER A

different laboratories can be facilitated. Even thisstep may not eliminate discrepancies because ofbatch to batch variation between the products ofthe same or different manufacturers. Our studieson different batches of MH agar from two manu-facturers showed reasonably good reproducibilityexcept with tetracycline, although statisticallysignificant differences were detected with severalother antibiotics. Comparison of different batchesfrom a single manufacturer, however, showed ahigh degree of agreement. This indicates that itshould be possible to obtain consistently repro-ducible media for susceptibility testing if criteriafor their preparation and their performance arecritically defined and disseminated. Recommenda-tions to this effect have been made elsewhere(Acta Pathol. Microbiol. Scand. Sect. B, Suppl.217, 1971), and Hoeprich et al. (8) have nowdescribed a fully synthetic medium which is satis-factory for susceptibility testing and opens up thepossibility of the use of such a medium for refer-ence purposes if not for day to day testing.A modification of susceptibility test media

which is frequently required involves the additionof blood to facilitate growth of more fastidiousorganisms. Since there are differences in theavailability of particular species of blood in differ-ent parts of the world, we investigated the effect ofcitrated and defibrinated horse, rabbit, sheep, andhuman bloods on susceptibility test results.Neither the species of blood nor the methods ofanticoagulation had any pronounced effect onresults except with tetracycline. With tetracycline,larger zone sizes were obtained with citratedblood, presumably owing to the binding of freecations. The extent of this effect with media otherthan that tested would depend on the concentra-tion of citrate used and the extent of tetracyclineinactivation by the base medium. Although anumber of other statistically significant differencesin zone diameters were encountered, they did notexceed 2 mm. It thus appears that defibrinatedblood from any of the mammalian speciesexamined can be considered interchangeable forroutine bacteriological purposes in susceptibilitytesting with the antibiotics examined.

It should be pointed out that the results pre-sented here were obtained with a limited numberof antibiotics and cannot be assumed to apply toall antimicrobial agents. Some comparisons with

SD SHERRIS ANTIMICROB. AG. CHEMOTHER*

sulfamethizole, for example, have shown signifi-cant variability between the products of the twomanufacturers of MH medium with a strain ofS. aureus (up to 4 mm) but none with a strain ofE. coli. Such observations reinforce the need forstandardization of media, at least by performancetests. The diffusion method used in these experi-ments differed somewhat from the referencemethod proposed by the International Collabora-tive Study and from the method of Bauer et al.(2). They were, however, sufficiently similar thatextrapolation of the conclusions appears justified.

LITERATURE CITED

1. Bauei, A. W. 1964. The significance of bacterial inhibitionzone diameters. A study of the factors influencing antibioticdisc sensitivity tests. Proc. Int. Congr. Chemother., 3rd,Stuttgart, p. 466-479.

2. Bauer, A. W., W. M. M. Kirby, J. C. Sherris, and M. Turck.1966. Antibiotic susceptibility testing by a standardizedsingle disc method. Amer. J. Clin. Pathol. 45:493-496.

3. Bulger, R. J., and K. Nielson. 1968. Effects of different mediaon in vitro studies of antibiotic combinations. Appl. Micro-biol. 16:890-895.

4. Ericsson, H. M., G. Tunevall, and K. Wickman. 1960. Thepaper disc method for determination of bacterial sensitivityto antibiotics. Scand. J. Clin. Lab. Invest. 12:408-413.

5. Garrod, L. P., and P. M. Waterworth. 1969. Effect of mediumcomposition on the apparent sensitivity of Pseudomonasaeruginosa to gentamicin. J. Clin. Pathol. 22:534-53g

6. Grove, D. C., and W. A. Randall. 1955. Assay methods ofantibiotics: a laboratory manual. Medical Encyclopedia,Inc.. New York.

7. Guthrie, F. 1958. A study of methods for determining sensitiv-ity of bacteria to antibiotics and the factors influencingdifferences in results. Amer. J. Med. Technol. 24:109-119.

8. Hoeprich, P. D., A. L. Barry, and G. D. Fay. 1971. Syntheticmedium for susceptibility testing. Antimicrob. Ag. Chemo-ther. 1970, p. 494-497.

9. Isenberg, H. D. 1964. A comparison of nationwide microbialsusceptibility testing using standardized discs. Health Lab.Sci. 1:185-256.

10. Kanazawa, Y. 1961. Clinical use of the disc sensitivity test.Antimicrob. Ag. Chemrnther. p. 926-942.

1 1. Kanazawa, Y. 1966. Single disc method for minimum inhibi-tory concentration (MIC) determination. J. Antibiot.(Tokyo) 19:175-189.

12. McFarland, J. 1907. The nephelometer. J. Amer. Med. Ass.49:1176-1178.

13. Neter, E., M. Murdock, and E. H. Kunz. 1952. The influenceof the composition of culture media on growth inhibition ofvarious microorganisms by antibiotic discs. Antibiot.Chemother. 2:35-41.

14. Trainer, T. D. 1964. The effect of the medium on the results oftesting sensitivity to antibiotics. Amer. J. Clin. Pathol.41:101-108.

15. Weinberg, E. D. 1957. The mutal effects of antimicrobial com-pounds and metallic cations. Bacteriol. Rev. 21:46-68.

O

on June 25, 2018 by guesthttp://aac.asm

.org/D

ownloaded from